Implementation of Feature Extraction and Improved Tracking Algorithm for Real Time Tracking of Multiple Moving Objects

Authors

  • Hyochang Ahn
  • Han-Jin Cho
  • June-Hwan Lee

Abstract

Recently, a real-time monitoring system needs to have the function of detecting and tracking the existence of an automatically moving object. These computer vision systems can be used in many fields because they can replace human roles. We propose a method to track objects by merging SURF, CAMShift, and Optical Flow. Through the SURF method, a search window is created for the region of interest for the object to be tracked, and the object is tracked through a combination of the CAMShift method and the optical flow. We also propose using Optical Flow to predict the position and velocity of a fast moving object. Experimental results show that the proposed method shows good results in terms of both accuracy and speed.

Keywords: CAMShift, Optical Flow, SURF, Object Detection, Object Tracking

Downloads

Published

2019-11-22

Issue

Section

Articles